Ah, yes. You can use sched_set_affinity to pin a process to a particular CPU, to answer both 1 and 2. I think its' supported in linux 2.6 and above, but I'm not totally certain about that. I've never done it myself but there are people on the list who have.
Lisa On Fri, Nov 20, 2009 at 2:43 PM, Amir Hossein Hormati < [email protected]> wrote: > 1- Is there anyway for me to understand which core(or cores) my benchmark > was running on? > 2- Do you think it's possible for the Linux scheduler to move the benchmark > from one processor to another while it's running? > > Thanks, > Amir > > > On Fri, Nov 20, 2009 at 5:41 PM, Amir Hossein Hormati > <[email protected]>wrote: > >> 1- Is there anyway for me to understand which core(or cores) my benchmark >> was running on? >> 2- Do you think it's possible for the linux scheduler to move the >> benchmark from one processor to another while it's running? >> >> Thanks, >> Amir >> >> >> >> On Fri, Nov 20, 2009 at 4:54 PM, Lisa Hsu <[email protected]> wrote: >> >>> Since you're running full system, it's probably because those processors >>> are in an idle loop. Check the user/kernel time stats. However, why 5/8 >>> would be high is unclear, I don't know why the last 3 wouldnt' also have a >>> high count.. >>> >>> Lisa >>> >>> On Fri, Nov 20, 2009 at 1:03 PM, Amir Hossein Hormati < >>> [email protected]> wrote: >>> >>>> Hi everyone, >>>> I setup a new benchmark by copying the related alpha binary to the disk >>>> image and setting up the related scripts. This benchmark is a single >>>> threaded benchmark which does some very trivial math operations and repeat >>>> for 1000 times. I ran m5 with 8 processors to see how this benchmark >>>> performs. After simulation was done, I looked at the stats file generated >>>> by >>>> m5. I expected to see one (or maybe two) cpu with high instruction count >>>> and >>>> the rest kind of idle, but out of the 8 processors 5 of them had very high >>>> instruction count. Can someone explain to me why this happened for a single >>>> thread program? >>>> >>>> Thanks, >>>> Amir >>>> >>>> _______________________________________________ >>>> m5-users mailing list >>>> [email protected] >>>> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users >>>> >>> >>> >>> _______________________________________________ >>> m5-users mailing list >>> [email protected] >>> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users >>> >> >> > > _______________________________________________ > m5-users mailing list > [email protected] > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users >
_______________________________________________ m5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
